rmuir commented on PR #12191: URL: https://github.com/apache/lucene/pull/12191#issuecomment-1456913229
My issue is still the concern about supporting such large dimension counts, it seems people want to just "use some off the shelf model" which is not necessarily geared at scalable search. But I think this is non-usecase, it never really gives good performance for your corpus right? I have a problem bumping limits without clear use-cases because we can't undo raising the limit really without pissing off users, as old indexes wouldn't be able to be read. So I'm still challenging the use-case for this. "because users ask for it" is not enough for me, given the maturity level of these technologies. When i look at the datasets for https://big-ann-benchmarks.com/, the most dimensions is 256. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
